Newswise —

Almost all of the Earth's inhabitants inhale air that surpasses the guidelines set by the World Health Organization (WHO). The situation worsens in metropolitan zones, which house over 50% of the world's population. To tackle the issue of air contamination, deemed the primary global environmental threat to health by the WHO, acquiring more dependable and precise information on air pollutant levels in our urban centers is vital, with a particular focus on nitrogen dioxide (NO2) due to its negative impact on quality of life and the consequential economic costs.

A group of scientists from the Earth System Services team of the Earth Sciences Department at the Barcelona Supercomputing Center - Centro Nacional de Supercomputación (BSC-CNS) has conducted a study to advance in this field of research. The study concludes that the use of artificial intelligence can be incredibly beneficial in obtaining dependable data on the likelihood of breaching legal thresholds for air pollution in urban areas. The objective of the study, published in the Geoscientific Model Development journal, is to aid in the enhancement of air quality management in cities by generating NO2 concentration maps at the street level every hour and quantifying the corresponding uncertainty.

For the first time, the new technique merges the outcomes of CALIOPE-Urban, an exceptional model in Spain that predicts air pollution at incredibly high resolutions of up to ten meters, at varying heights, and at any location in the city, with a comprehensive urban database that encompasses official air quality station observations, low-cost sensor campaigns, details on building density, meteorological factors, and a vast range of other geospatial data. Thus, it is possible to pinpoint regions in the city where the current monitoring system requires improvement, assisting in the streamlining of plans to minimize air pollution.

Jan Mateu, the leader of the BSC Air Quality Services team and one of the primary authors of the research, states that "the application of artificial intelligence in combination with CALIOPE-Urban predictions and the integration of all urban data help enhance the model. When the simulation fails to explain the spatial pattern of pollution, we can implement machine learning to rectify and enhance the forecast."

By utilizing machine learning methods with observational data gathered from previous campaigns that employed passive dosimeters, the study represents a significant leap forward in reducing the inherent uncertainties linked to air quality models, caused by the scarcity of monitoring stations. This technique enables a more accurate spatial analysis of excessive air pollution in various areas of the city.

The research, which concentrated on the city of Barcelona during the pilot phase, established that the Eixample district has the poorest air quality in the Catalan capital. The study concludes that 95% of this region has over a 50% probability of exceeding the yearly average NO2 threshold of 40 μg/m3 as defined by the European Commission (European Air Quality Directive 2008/50/EC).

Álvaro Criado, one of the primary authors of the study and a researcher in BSC's Air Quality Services team, notes that "The Eixample district, which is the most populous area in Barcelona, is the most impacted region in the city, with the vast majority of its surface area having more than a 50% likelihood of surpassing the yearly NO2 limit established by the European Commission. Our methodology will enable the public administration to design and implement policies to enhance air quality in urban regions, which is particularly important since air pollution constitutes the primary environmental risk factor for human health."

The CALIOPE-Urban model

CALIOPE-Urban, developed by the BSC, is a modelling tool that evaluates the concentration of nitrogen dioxide (NO2) at street level in Barcelona; however, it could be extended to other cities or metropolitan regions. NO2 and its precursors primarily arise from combustion sources, such as vehicle engines, making monitoring critical in large cities where traffic is typically congested to fight air pollution.

The system is distinct in Spain and provides valuable information to citizens and air quality managers on how traffic impacts air pollution in every neighborhood. This data is critical for designing and implementing effective planning and mitigation strategies to safeguard citizens from the health hazards of air pollution. Currently, CALIOPE-Urban is directed towards Barcelona, but efforts are already underway to extend it to other municipalities in partnership with several municipal and regional administrations.

CALIOPE-Urban is a combination of two technologies, the CALIOPE regional model, and an urban model that considers street-level air pollution, traffic emissions, and meteorological data. CALIOPE is the only air quality prediction system that offers operational forecasts for Barcelona, Catalonia, the Iberian Peninsula, and Europe. It is also the only Spanish contributor to the European Union's Copernicus Atmosphere Monitoring Service (CAMS).

Journal Link: Geoscientific Model Development